Skip to content

Conversation

@PeterDaveHello
Copy link
Contributor

@PeterDaveHello PeterDaveHello commented Apr 4, 2025

This is a follow-up to #2233, addressing some missing and newly found improvements from the real-use scenario with enough context. I believe this will really help make the zh-TW Traditional Chinese locale sound more natural and professional and less machine- or AI-translated (as I guess this is how the original version came about).


Summary from GitHub Copilot for your reference:

This pull request includes several changes to the Traditional Chinese (zh-TW) localization files to improve translations and consistency across the project. The main updates focus on refining terminology, enhancing readability, and ensuring uniformity in the translated content.

Localization improvements:


Important

Improves zh-TW Traditional Chinese localization by refining translations for clarity and consistency across multiple files.

  • Localization Improvements:
    • README.md: Updated terms for clarity, e.g., "自定義" to "自訂", "文件" to "檔案".
    • common.json: Enhanced translation accuracy for confirmation, error, and warning messages.
    • tools.json: Improved translation for line range descriptions in readFile tool.
    • chat.json: Refined translations for directory operation messages.

This description was created by Ellipsis for 9cb3aa8. It will automatically update as commits are pushed.

Continuously improve the translation for a better quality.
@changeset-bot
Copy link

changeset-bot bot commented Apr 4, 2025

⚠️ No Changeset found

Latest commit: 9cb3aa8

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

This PR includes no changesets

When changesets are added to this PR, you'll see the packages that this PR includes changesets for and the associated semver types

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR

@dosubot dosubot bot added size:L This PR changes 100-499 lines, ignoring generated files. enhancement New feature or request labels Apr 4, 2025
Copy link
Collaborator

@mrubens mrubens left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thank you!

@dosubot dosubot bot added the lgtm This PR has been approved by a maintainer label Apr 4, 2025
@mrubens mrubens merged commit 359a2f5 into RooCodeInc:main Apr 5, 2025
21 checks passed
@github-project-automation github-project-automation bot moved this from New to Done in Roo Code Roadmap Apr 5, 2025
@mrubens
Copy link
Collaborator

mrubens commented Apr 5, 2025

@PeterDaveHello I asked Roo to update its translation instructions based on the changes in this PR. Do these look like they're on the right track?

Screenshot 2025-04-04 at 10 41 19 PM

@PeterDaveHello PeterDaveHello deleted the zh-TW branch April 5, 2025 10:19
@PeterDaveHello
Copy link
Contributor Author

@mrubens The direction looks fine, but there's a fundamental challenge: Given that Chinese content is already a minority in global internet training data, and Traditional Chinese data is significantly less available compared to Simplified Chinese, this naturally becomes a weakness for LLMs. This is likely a common challenge for many non-mainstream languages - while LLM translations might be generally comprehensible, they often lack naturalness and clarity in certain areas. Of course, we all acknowledge this is already a substantial improvement over traditional machine translation.

From my experience, while frontier models like GPT-4.5, Claude 3.5, Gemini 1.5 Pro, and Mistral Large 2, Grok 3 show relatively better performance in Traditional Chinese, they still require extensive prompt engineering, particularly to avoid confusion with some Simplified Chinese terms.

If we could integrate this workflow into the project, I'd be happy to help optimize the prompts. Meanwhile, I'll continue providing manual translations and revisions periodically. One potential improvement would be instructing LLMs to reference existing human-edited translations and maintain consistency with the English source, which might yield better results.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

enhancement New feature or request lgtm This PR has been approved by a maintainer size:L This PR changes 100-499 lines, ignoring generated files.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants